It’s more discussed here and i have some G80 DCP profiles which are remade(v2) for the new DxOPL.
(not me some one else i just use them.)
The V1 where having some issues in saturation in orange and the clipping highlight tool of DxO was blinking a lot. (clipping tool is not only lumination (brightness) but also saturation.)
So he was recalibrating them for dxoplv3.
I can’t find the text but i assume he uses every data he can squees out of the sensor RAW data.
If i compare the Generic one of DxO and the Huelightv2 versions the clipping warning is often active.
Can someone explain to me how i have to interpreted this DCP profiling?
Does the generic renderings are a better choise because the don’t clip so much or is the clipping just in about 245 and real border is 250 so generic leaves a lot of data out the image?
i can see that it pushes more to the edges in the histogram but the why and theoretical background i don’t know.
i know that rawdigger? and fastrawviewer show much more then a dxo Histogram can be showing so what do i have to think about this behaviour?
What impact does using the “protect saturated colours” have on what you are seeing?
As far as i know it suppresses the spikes in the rgb channels at the right side of the histgram.
not the lumination (brightnes combined rgb) but more the single channel. So it helps to supress the out of gamut colors i think.
The highlight blinkies (sun) is also a oversaturation on the bright side warning. (if you see the lumination still under 245 but one of the rgb channels is around 245-250 you see it react. Same working of the moon, in the darker area’s.
i find DxO highlight blinkies earlier reacting then in other applications. That’s why i try to understand the dcp. it’s developed for adobes LR and ACR. DxO’s own generic rendering is less “agresif”? less deap black and bright white. less saturated colors.
I have a yellow fish which shows fairly the difference in bright “yellow”.
The usual idea of dcp profiles is to provide means to reproduce colours as faithfully as possible.
When a colour is outside the gamut as shown in your tool, the image is overexposed or the tool’s working colour space is too small. Tweaking the profile denies the original meaning (faithfulness) of using the profile.
Instead of tweaking profiles, you could also just desaturate or darken the offending colour with the hsl tool. Or redo the shot with correct exposure and possibly set AdobeRGB in camera.
Rendering is interpreting data from the file in order to get a (possibly pleasing, correct, good…) image. Using different profiles provides different interpretations. Chose the profile that delivers an image you like and you are set. Whether coulours clip or not is not really that important if you like what you get and unless you need absolute faithfulness in a reproduction.
“Good” and “better” can only be used within the intended scope which can be either “like what I get” or “absolutely faithful” or anything else.
And then, there is printing, which introduces fairly different colour spaces as shown in this example: https://de.wikipedia.org/wiki/Farbraum#/media/Datei:Gamut_ISOcoated_and_sRGB.png
As far as i understand a color profile is used to crunch the numbers stored in the rawfile representing channels RGB to actual colors you can see.
( it’s a difficult subject and thinking in dutch translating to englisch isn’t make it easier to explain thoughts/discus. )
1 when i have more time i wil set up a case which shows generic , dcp v172 version 1. And the new version v2 for my g80 , examples.
This wil makes my question clearer.
2 camera type/brand depending colorprofiles are mostly for getting a nicer realist look out the sensors data. Most sensors has more exposure tolerance for one colorrange/ hue then over the other.RGB channels arn’t equally working so to speak. By tweaking that you seem to balance this better creating a more faitefull look/ image.
(at least that my knowledge or thinking how colorprofiling works.)
In this matter i have standard, medium, high, low , portrait. So the sensor readout is preset changed to create a contrast rich (saturation) or more centre balanced preset.
My Canon Cameras have a few preset “picture styles” that will produce distinctly different results when they are applied to whatever photo I use. These picture styles define how data from r, g and b sensor pixels are combined. Same thing is true for profiles.
Whatever picture style or profile I use only matters if I need to produce faithful reproductions. The things that matter in general photography is if you like what you get or if your clients like what you create for them.
I think of different styles and profiles as recipes that allow me to cook the food I like. As with cooking, it helps to stick to a recipe the first few times. After a while, variations enrich your capabilities as a cook and help you create wonderful food from scratch. Still, microwaving a frozen pizza might be something you like. Is it right or wrong? Politicians try to sell you right and wrong, as photographers, we can also deal with things in between…